aiqbench analysis
This research advances robotic manipulation by using generative video models to synthesize bare-hand visual data from gloved recordings, effectively bridging the simulation-to-reality gap for tactile force prediction. By recovering dense contact signals purely from egocentric vision, TouchSight reduces the need for intrusive hardware while improving generalization across unseen datasets. The core innovation lies in leveraging generative AI to augment training data, enabling robust tactile inference without physical instrumentation at capture time.
TouchSight: Bare-Handed Tactile Prediction from Egocentric Video via Generative Visual Augmentation
arXiv:2609.20414v1 Announce Type: cross Abstract: Tactile signals provide direct contact and force measurements that are essential for understanding physical interactions and enabling dexterous robotic manipulation. However, tactile sensing requires direct measurement at contact interfaces, making large-scale data collection reliant on intrusive, costly, and restrictive instrumentation. We present TouchSight, a monocular egocentric vision framework for dense full-hand contact force prediction that leverages 500 hours of pressure-glove recordings and extensive hand-object interaction (HOI) data. To address the appearance gap between gloved training data and bare-hand real-world scenarios, we construct TwinTouch-20H: 20 hours of paired visual data in which generative video models re-render gloved recordings as bare-hand observations against new backgrounds while preserving the original measured tactile labels. TouchSight predicts dense force from both gloved and generated bare-hand videos, outperforms prior contact prediction methods on OakInk2, qualitatively generalizes to natural bare-hand egocentric videos from unseen datasets, and improves consistently as glove supervision scales. These results demonstrate that dense tactile signals can be recovered from egocentric vision alone, without tactile instrumentation at capture time.
Abstract:Tactile signals provide direct contact and force measurements that are essential for understanding physical interactions and enabling dexterous robotic manipulation. However, tactile sensing requires direct measurement at contact interfaces, making large-scale data collection reliant on intrusive, costly, and restrictive instrumentation. We present TouchSight, a monocular egocentric vision framework for dense full-hand contact force prediction that leverages 500 hours of pressure-glove recordings and extensive hand-object interaction (HOI) data. To address the appearance gap between gloved training data and bare-hand real-world scenarios, we construct TwinTouch-20H: 20 hours of paired visual data in which generative video models re-render gloved recordings as bare-hand observations against new backgrounds while preserving the original measured tactile labels. TouchSight predicts dense force from both gloved and generated bare-hand videos, outperforms prior contact prediction methods on OakInk2, qualitatively generalizes to natural bare-hand egocentric videos from unseen datasets, and improves consistently as glove supervision scales. These results demonstrate that dense tactile signals can be recovered from egocentric vision alone, without tactile instrumentation at capture time.
Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
arXiv:2609.20414 [cs.CV]
(or arXiv:2609.20414v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2609.20414 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 265 words
Related coverage
Dyna-2: General Purpose Physical Agents with Hour-Long Autonomy — Jason Ma, Dyna Robotics
Latent Space · Oct 13, 2026
ExoBridge: Learning a Bare Hand to Hand-Worn Exoskeleton Mapping through Human Limb Coupling
arXiv · Oct 7, 2026
Dexterous Tactile World Model
arXiv · Sep 30, 2026
RLHND: Video Foundation Models as Physically Grounded Hand Trackers for Robot Learning
arXiv · Oct 8, 2026